Differential Roles of Fibrinogen and von Willebrand Factor on Clot Formation and Platelet Adhesion in Reconstituted and Immune Thrombocytopenia

Abstract

BACKGROUND: 

Bleeding tendencies in immune thrombocytopenia (ITP) do not always correlate with the number of platelets, suggesting platelet function variation. We used a model of normal whole blood thrombocytopenia to compare platelet function and other hemostatic variables with ITP patients. We further investigated the effect of in vitro spiking with von Willebrand factor (vWF) and fibrinogen on platelet function and hemostatic variables.

METHODS: 

The Cone and Plate(let) Analyzer was used to measure platelet adhesion (surface coverage [SC], %) and aggregation (average size, μm²) under defined shear rate (1200 s⁻¹). Rotational thromboelastometry was used to determine variables of clot formation triggered by CaCl₂ and tissue factor.

RESULTS: 

In both the model of thrombocytopenia as well as in ITP, the SC and to some extent the average size were correlated to the platelet number over a range of 5 to 80 × 10⁶/mL. The results obtained for most ITP samples were within the boundaries of the lower and upper limits set by the whole blood model of thrombocytopenia. The addition of 2 U/mL vWF (Haemate-P) to whole blood (calculated to plasma volume) results in an increase in the SC and average size without affecting clot formation. Spiking with fibrinogen (100 and 300 mg/dL) did not affect platelet deposition but improved clot formation.

CONCLUSIONS: 

Using a model of whole blood thrombocytopenia enables us to establish reference variables for the Cone and Plate(let) Analyzer and rotational thromboelastometry and to assess platelet function and clot formation in the presence of severe thrombocytopenia. We demonstrated that in most cases of ITP, platelet function is comparable to normal platelets. This work also suggests that vWF and fibrinogen differentially affect primary and secondary hemostasis and therefore both may perform a function in the bleeding phenotype and possibly may be considered for treatment in patients with ITP.

In Brief

Published ahead of print April 7, 2011

Immune thrombocytopenia (ITP) is an autoimmune disorder that is characterized by low platelet count due to accelerated platelet destruction as a result of autoantibodies against platelet surface antigens and T-cell toxicity.^1–3 Occasionally, impaired megakaryocytopoiesis has been found.^4,5 The consequence of thrombocytopenia (TCP) is complex because platelets are important not only in building up the primary hemostatic plug but also for thrombin generation and clot retraction.^6,7 Bleeding symptoms in ITP are variable, ranging from mild to severe and occasionally may even be life-threatening. The conventional way to manage patients according to the degree of TCP does not include the assessment of platelet function. Bleeding is usually associated with a reduction of platelet count below 20 × 10⁶/mL. However, in some cases, bleeding is absent even at platelet counts of 10 × 10⁶/mL or lower. Conversely, bleeding may be present at platelet counts that exceed 30 × 10⁶/mL. These clinical observations suggest the presence of platelet function diversity and that more defined ways to evaluate platelet function as well as strategies to improve management in case of severe bleeding are needed. Current expert consensus for first-line treatments includes the use of steroids, anti-D, and immune globulins. A response can take several days to weeks depending on the type of treatment, and side effects may be severe.⁸ Also, a justified use of these treatments before invasive procedures is based on expert recommendations in regard to the platelet numbers only.⁹

For the purpose of platelet function evaluation, neither platelet aggregation studies nor standard clotting assays may serve for comprehensive platelet function and clot formation assessment in ITP. Of note, light transmittance platelet aggregometry, the “gold standard” method of platelet function assessment, cannot be used when the platelet count is <100 × 10⁶/mL.¹⁰ Moreover, platelet deposition in whole blood (WB) under defined shear stress is much more relevant for platelet studies approximate to blood flow in arterial vessel walls. The Cone and Plate(let) Analyzer (CPA; DiaMed Inc., Crefier, Switzerland) has been developed and has proved to be useful for the assessment of platelet function and for monitoring of antiplatelet and replacement therapy.^11–14

In the last decade, the rotational thromboelastometry (ROTEM) has emerged as a valuable point-of-care device and a research tool for the assessment and monitoring of hemostasis and fibrinolysis.^15,16 In that respect, ROTEM has been used in different experimental models including TCP to evaluate hemostasis in both plasma and WB.^7,17–20

Using a model of WB is obviously advantageous because it integrates the contribution of blood cells (platelets, leukocytes, erythrocytes, as well as plasma components) with the process of hemostasis.^18,19,21 However, with WB models, a platelet count <20 × 10⁶/mL is difficult to achieve. In this study, we used an original WB TCP model ranging from as low as 5 × 10⁶ and up to 80 × 10⁶ platelets per mL. The model was used to establish normal values for platelet adhesion/aggregation function measured with the CPA and clot formation (ROTEM) at different platelet counts and to compare them with platelets from ITP patients. This approach will be able to link between platelet counts, platelet function, and blood clotting in ITP patients and possibly in patients with other forms of TCP.

METHODS

Blood Samples

The study was approved by the local ethics committee and written informed consent was obtained from all participants. Blood samples were drawn from healthy subjects with normal platelet counts and from patients with chronic ITP who were diagnosed according to the international consensus guidelines.²² None of the subjects received drugs affecting platelet function or the coagulation system within the last 14 days. Blood samples for cell count were collected in 2.7-mL tubes containing 1.6 mg/mL EDTA. Blood samples for CPA and ROTEM analysis were collected in 3.8-mL tubes containing 0.38 mL buffered sodium citrate (0.106 M). Platelet numbers were measured in the Beckman Coulter LH750 (Beckman Coulter, Inc.). von Willebrand factor (vWF) activity was measured as ristocetin cofactor activity using the platelet aggregometer PACS 4 (Helena Laboratory, Beaumont, TX). It ranged in both healthy subjects and ITP patients from 70% to 130%.

Reconstitution of WB TCP

Platelet-rich plasma (PRP) was prepared by centrifugation of WB at 180g for 12 minutes. Packed cells (PCs) and platelet-poor plasma (PPP) were prepared by centrifugation of the remaining blood at 1200g for 12 minutes. For WB reconstitution of platelets at numbers <20 × 10⁶/mL, PPP was additionally subjected to high-speed centrifugation (10,000 rpm) for 3 minutes and the red cells were aspirated from the lower layer of the PCs. PRP and PC samples were counted and WB samples were reconstituted at various platelet levels according to the following formulation:

where x = hematocrit in WB/hematocrit in PCs and y = desirable platelet number in WB/platelet number in PRP. x, y, and z represent the ratio of blood components volume needed for reconstitution of WB. WB samples were prepared at the following platelet levels: 5, 10, 20, 40, 80, and 160 × 10⁶/mL.

Cone and Plate(let) Analyzer

In CPA, sodium citrate anticoagulated blood samples (130 μL) are placed on polystyrene wells (Nunc, Roskilde, Denmark) and subjected to flow at 1200 s⁻¹ for 2 minutes using a specially designed conical disk. Fibrinogen and vWF interact with the polystyrene surface mediating platelet adhesion.²³ Wells were thoroughly washed, stained with May-Grünwald stain, and analyzed using an image analyzer (DiaMed, Cressier, Switzerland). Both glycoproteins Ib/IX and IIb/IIIa are needed for platelet deposition in this system.¹³ Platelet adhesion was determined by measuring the percentage of the well surface that was covered with platelets (surface coverage [SC]), and the platelet aggregation on the surface was determined by measuring average size (AS) of the adhered particles in μm².

Rotational Thromboelastometry

Variables of clot dynamics were measured by the ROTEM (Pentapharm, Munich, Germany). Clot formation was triggered by CaCl₂ (final concentration, 16.6 mM) and tissue factor. (The EXTEM reagent is a ready-to-use thromboplastin from rabbit brains provided by Pentapharm.) Tests were performed at 37°C in 4 channels simultaneously. The following ROTEM parameters were used to evaluate hemostasis: clotting time (CT), clot formation time (CFT), and maximum clot firmness (MCF). CT is a period from measurement start until start of clot formation and reflects the time needed for thrombin generation. CFT reflects the initial phase of clot formation. MCF is the maximum firmness that the clot has reached. CT was used to ensure that vWF and fibrinogen did not change this parameter. CFT and MCF were expected to be the most informative variables after spiking with fibrinogen and probably with vWF. Normal reference values of these variables were published elsewhere and confirmed in our preliminary assays.²⁴ The reconstituted WB TCP as well as CPA and ROTEM experiments were performed by the same operator.

Spiking Experiments with Fibrinogen and vWF Concentrates

Fibrinogen concentrate (Haemocomplettan® P; CSL Behring, King of Prussia, PA) and vWF (Haemate®-P; Aventis Behring GmbH, Marburg, Germany) were used separately and in combination to assess their ability to enhance platelet function and hemostasis. In both the CPA and ROTEM experiments, samples were spiked with fibrinogen at concentrations of 100 and 300 mg/dL and with vWF at a concentration of 2 U/mL.

Statistical Analysis

Results are shown as the interquartile range (25th–75th percentiles), median, and whiskers on the boxes corresponding to the 5th/95th percentiles. The data were processed using the Friedman test followed by a post hoc Dunn test. Each set of groups corresponding to a distinct platelet count was analyzed separately. The best fit curves to the platelet concentration response in the WB model of reconstituted TCP were derived by nonlinear regression analysis, and the 95% prediction bands were established (Fig. 1). The area included both the uncertainty in the true position of the curve and also accounted for scatter of data around the curve. In ITP patients, correlation between platelet count and hemostasis parameters was evaluated using the Pearson coefficient (r).

Figure 1: A, Surface coverage (SC), and B, maximum clot firmness (MCF) as a function of platelet count in reconstituted normal blood and in blood samples from immune thrombocytopenia (ITP) patients. The 95% prediction bands of the hemostasis parameters in reconstituted normal blood samples are enclosed within the dashed lines. The solid lines represent the best fit curves to the platelet concentration-response relationship derived by nonlinear regression analysis. The values of SC obtained for ITP patients are presented with open circles (control) and solid circles (addition of 2 U/mL von Willebrand factor in Cone and Plate(let) Analyzer). MCF values are presented with open circles (control) and solid circles (addition of 300 mg/dL fibrinogen in rotational thromboelastometry).

RESULTS

Blood Samples

The study enrolled 16 healthy subjects (12 men and 4 women) and 12 adult patients with chronic ITP (8 men and 4 women) with a matched age ranging from 20 to 87 years.

Platelet numbers in healthy subjects were between 160 and 230 × 10⁶/mL and in ITP patients from 2 to 74 × 10⁶/mL. There was no difference in fibrinogen level between the 2 groups ranging from 240 to 360 mg/dL in the normal subjects and from 220 to 370 mg/dL in ITP patients.

CPA Analysis of Platelet Function in a Model of WB Reconstituted TCP

Reconstitution of normal WB to achieve platelet counts of 5, 10, 20, 40, 80, and 160 × 10⁶/mL was performed as described in the Methods section. Blood samples were divided into 4 separate tubes: controls, samples spiked with 100/300 mg/dL fibrinogen, 2 U/mL vWF, and both reagents (the amount of the reagents was calculated to plasma volume in the WB). After 20 minutes of incubation, the samples were placed in polystyrene wells subjected to a shear rate of 1200 s⁻¹ for 2 minutes.

Results are shown in Figure 2. Both SC and AS positively correlate to the platelet numbers. Median values for SC increased from 1.3% to 11% and AS from 25.4 to 43 μm² corresponding to an increase in platelet counts from 5 to 160 × 10⁶/mL. Spiking with vWF results in a 1.6- to 1.8-fold increase in SC over all ranges of platelet counts and 1.2- to 1.3-fold increase in AS only at a platelet count of ≥10 × 10⁶/mL. Fibrinogen had no effect on either the SC or AS nor did it have any additive effect to vWF.

Figure 2: Platelet deposition in whole blood under shear condition (1200 s⁻¹) with reconstituted thrombocytopenia. Normal blood was subjected to differential centrifugation to obtain platelet-rich plasma, platelet-poor plasma, and packed cells. Whole blood samples were reconstituted by mixing these blood components with different platelet levels. The blood samples were supplemented with phosphate-buffered saline (control), fibrinogen (300 mg/dL), von Willebrand factor (vWF) (2 U/mL), or fibrinogen plus vWF. After 20 minutes of incubation, the samples were assayed in a Cone and Plate(let) Analyzer device. The medians and interquartile ranges (25th, 75th percentiles) are shown. The whiskers on the boxes correspond to the 5th/95th percentiles. Data of surface coverage (%) and average size (μm²) of adhered platelets (A and B) were processed by using the Friedman test followed by a post hoc Dunn test. *P < 0.05, **P < 0.01, ***P < 0.001 (n = 8–12).

Clot Formation in a Model of WB Reconstituted TCP

WB reconstituted TCP was prepared as described. No difference was observed in CFT and MCF between the blood samples with a platelet concentration of 5 and 10 × 10⁶/mL (Fig. 3). However, with platelet numbers between 10 and 160 × 10⁶/mL, CFT gradually decreased and MCF gradually increased (457 to 108 seconds and 30.5 to 58 mm, respectively). There was no association between CT and platelet number (data not shown). Spiking with 300 mg% fibrinogen results in a 2- to 3.4-fold reduction in CFT and in a 1.2- to 1.3-fold increase in MCF.

Figure 3: Normal whole blood clotting in reconstituted thrombocytopenia. Normal blood was subjected to differential centrifugation to obtain platelet-rich plasma, platelet-poor plasma, and packed cells. Whole blood samples were reconstituted by mixing these blood components with different platelet levels. The blood samples were supplemented with phosphate-buffered saline (control), fibrinogen (300 mg/dL), von Willebrand factor (vWF) (2 U/mL) or fibrinogen plus vWF. After resting for 20 minutes, the samples were assayed in a rotational thromboelastometry device. Blood clotting variables were determined using EXTEM, i.e., by addition of CaCl₂ and tissue factor. The medians and interquartile ranges (25th, 75th percentiles) are shown. The whiskers on the boxes correspond to the 5th/95th percentiles. Clot formation time (seconds) (A) and maximum clot firmness (mm) (B) data were processed by using the Friedman test followed by a post hoc Dunn test. *P < 0.05, **P < 0.01, ***P < 0.001 (n = 8–12).

Spiking with vWF at a concentration of 2 U/mL as single agent or with fibrinogen had no effect on any of the clot formation variables.

CPA and ROTEM Analysis of Blood Samples from ITP Patients

Twelve samples from patients with chronic ITP were assayed. Platelet counts ranged from 2 to 74 × 10⁶/mL. In all except 2 patients, SC demonstrated positive correlation to platelet counts in control platelets (Table 1). In 1 sample with a platelet count of 47 × 10⁶/mL, the SC was lower and in another sample with a platelet count of 67 × 10⁶/mL, it was higher than expected. Overall, both SC and AS values correlated to the platelet counts as determined by the Pearson coefficient (for SC: r = 0.81, P = 0.015; for AS: r = 0.61, P = 0.037). Also, both SC and AS values were increased after spiking with vWF (median values increase from 2.8% and 34 μm² in intact samples to 7.3% and 47.5 μm² after spiking, respectively). In contrast, spiking with fibrinogen (either 100 or 300 mg/dL) had no effect on platelet function and no additional value to vWF alone.

Table 1: Effect of Fibrinogen, von Willebrand Factor, and Both Agents on Platelet Function Under Flow Conditions in Immune Thrombocytopenia Patients

In the ROTEM/EXTEM test, CFT had shown negative correlation (r = −0.86, P = 0.0007) whereas MCF had shown positive correlation (r = 0.89, P = 0.0007) to the platelet counts (Table 2). Both CFT and MCF values measured in samples of ITP patients improved by spiking with 300 mg/dL fibrinogen but not with 100 mg/dL (CFT decreased from 198 to 138 seconds, P < 0.001; MCF increased from 45 to 51 mm, P < 0.01). CT was not affected by fibrinogen. Similar to WB, reconstituted TCP spiking with vWF had no effect on any of the clot formation variables nor did it have any additive effect to fibrinogen.

Table 2: Effect of Fibrinogen, von Willebrand Factor, and Both Agents on Clot Dynamics in Immune Thrombocytopenia Patients

Comparison of the Results Obtained from ITP Patients with WB Reconstituted TCP

The 95% prediction bands were constructed based on the values of primary and secondary hemostasis obtained from the WB reconstituted TCP at platelet levels from 5 to 160 × 10⁶/mL. SC was chosen as an integral parameter expressing both platelet adhesion and aggregation. In all except 2 ITP patients, SC decreased into the range of the normal reconstituted blood suggesting normal platelet function (open circles in Fig. 1A). After spiking with vWF, the function of normal platelets from reconstituted WB as well as platelets from ITP was improved (solid circles in Fig. 1A). In 1 patient, SC was higher than the upper level of the “normal” range with a further increase by addition of vWF.

Regarding ROTEM analysis, MCF was chosen for the construction of the 95% predictive bands because it is generally dependent on both coagulation and platelet activity. In all ITP patients, MCF decreased into the range of the normal reconstituted blood. MCF further increased after spiking with 300 mg/dL fibrinogen. In 1 ITP sample, after spiking, MCF was even higher compared with the level of reconstituted blood with the same platelet count (Fig. 1B).

DISCUSSION

The incidence of major or fatal bleeding in ITP patients is relatively low and usually occurs at platelet counts below 20 × 10⁶/mL.^25,26 Treatment must be tailored for each patient according to the estimated bleeding risk, which varies with lifestyle, comorbid conditions, and the need for invasive procedures. First-line treatments include the use of corticosteroids, anti-D, and immune globulins and the response can take several days to weeks depending on the type of treatment, and side effects may be severe.⁸ A justified use of these treatments before invasive procedures is based on expert recommendations in regard to the platelet numbers only and without considering platelet function and WB clot dynamics.⁹ The assessment of platelet function in ITP patients as in all other patients with TCP is not easy, especially because the gold standard method of platelet function, the optical platelet aggregometry, cannot be used when the platelet count in PRP is <100 × 10⁶/mL. Fluorescence flow cytometry is a reliable test for platelet activity and can be used even in the presence of substantial TCP. However, Panzer et al.²⁷ evaluated P-selectin expression on platelets and found that it neither correlates with platelet counts nor to the bleeding score in ITP patients. In the same study, only platelet adhesion as measured with the CPA correlated with bleeding symptoms and platelet count. Taking these data one step forward, we created a model of WB reconstituted TCP at different platelet levels to establish “normal” reference values for platelet adhesion/ aggregation and for ROTEM parameters at a range of platelets between 5 and 160 × 10⁶/mL. The boundaries between 5th to 95th percentile were determined, and platelet function and clot dynamics in patients with ITP could be evaluated.

The results, which are based on the samples tested, prove that in most ITP patients, the platelets display normal function in both adhesion/aggregation assays (CPA) and clot dynamic assays (ROTEM). That is to say that bleeding tendency is primarily dependent on platelet numbers and is less attributable to platelet malfunction. Therefore, increasing the platelet number would be the mainstay treatment in bleeding situations. However, as mentioned above, first-line treatment modalities have their limitations and risks so a different approach to improve hemostasis would be helpful. The important role of vWF in platelet adhesion under high shear stress is well documented.^28,29 Recent evidence suggests that vWF also protects fibrinogen against degradation by plasmin.³⁰ Thus, in clinical situations with increased fibrinolysis, vWF may help to preserve the adhesive role of fibrinogen in platelet-rich thrombi. Indeed, we have demonstrated that in the model of WB TCP, the introduction of vWF at a concentration of 2 U/mL significantly improved platelet adhesion under shear stress but had no effect on clot dynamics as measured by ROTEM. The same results were obtained with platelets from ITP patients. Fibrinogen is useful at enhancing clot formation in different models of coagulopathy including the presence of TCP.^31–35 The ability of fibrinogen to compensate for profound TCP, even better than platelet replacement, was also demonstrated in vivo using an animal model of severe trauma.¹⁵

In this study, spiking with fibrinogen at a dose of 300 mg/dL (but not 100 mg/dL) was followed by a platelet number–dependent decrease in CFT and increase in MCF. This is in agreement with the findings that in vitro spiking with fibrinogen at a concentration no less than 200 mg/dL can optimize the rate of clot formation.^33,34 However, only small amounts of fibrinogen are needed as bridging ligands to the glycoprotein IIb/IIIa to ensure platelet adhesion. This is why platelet deposition in the CPA system was unaffected by spiking with a high concentration of fibrinogen. As with the CPA, platelets from ITP patients displayed comparable function in ROTEM to normal platelets. The results should be interpreted in light of the differences and limitations of both methods. In the CPA, anticoagulated blood is tested and therefore platelet function is evaluated with no involvement of the coagulation system. With ROTEM, both the coagulation system and platelet function are performing a role but without the effect of blood flow. However, using both methods may improve the preoperative hemostatic assessment in ITP patients.

In summary, the combined measurement of adhesion/ aggregation and clot dynamics parameters enable a more comprehensive view of the hemostatic process. This study demonstrates that vWF improves platelet adhesion and that fibrinogen increases the viscoelasticity of a clot in ITP blood samples in a manner similar to TCP reconstituted with normal WB. The utility of this approach in the routine work-up of patients with ITP or other clinical scenarios with TCP, before performing an invasive procedure, should be tested in prospective clinical studies.

DISCLOSURES

Name: Mudi Misgav, MD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Mudi Misgav has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Boris Shenkman, MD, PhD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Boris Shenkman has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Ivan Budnik, MD.

Contribution: This author helped design the study and analyze the data.

Attestation: Ivan Budnik has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Yulia Einav, PhD.

Contribution: This author helped design the study and analyze the data.

Attestation: Yulia Einav has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.

Name: Uri Martinowitz, MD.

Contribution: This author helped design the study, analyze the data, and write the manuscript.

Attestation: Uri Martinowitz has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.